ITE Technical Report
Online ISSN : 2424-1970
Print ISSN : 1342-6893
ISSN-L : 1342-6893
37.12
Showing 1-21 articles out of 21 articles from the selected issue
  • Type: Cover
    Pages Cover1-
    Published: February 26, 2013
    Released: September 21, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Download PDF (11K)
  • Type: Index
    Pages Toc1-
    Published: February 26, 2013
    Released: September 21, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Download PDF (59K)
  • Hotaka WATARAI, Kiyoshi HOSHINO
    Type: Article
    Session ID: HI2013-41/3DIT2013-9
    Published: February 26, 2013
    Released: September 21, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    We propose a virtual keyboard without installing anything on your hands, make a character input by user's gesture. We narrow down keys can be entered using the position of the center of gravity of the hand. We detect a typing operation using fingertip positional change in rectangular frame set to each finger. Combination of these, we construct virtual keyboard that can input keys even if hands move in image. We experimented with this system. As a result, skilled person can enter keys high percentage of correct and stable input time. We confirmed non-expert gets used to this system in short time, input time and percentage of correct is improved.
    Download PDF (9800K)
  • Noriaki FUJISHIMA, Kiyoshi HOSHINO
    Type: Article
    Session ID: HI2013-42/3DIT2013-1
    Published: February 26, 2013
    Released: September 21, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Previous fingernail detection method was not good at removing some skin areas which have nail-color and gloss. We found that color continuity in skin area can be used for removing these areas. In this paper, therefore, we propose new fingernail detection method using distribution density of strong nail-color pixels and color continuity. We investigated relations between wrist rotation angles and percentage of correct detection. The number of users was three. As a results, we confirmed that our propose method could detect only fingernails at least 85% from ±40 to ±90 and raise accuracy compared with previous method.
    Download PDF (7075K)
  • Naoya TSUSHIMA, Masashi KAMEDA
    Type: Article
    Session ID: HI2013-43/3DIT2013-1
    Published: February 26, 2013
    Released: September 21, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    The positive/negative sign code of the DCT coefficients is important to reconstruct the structural information of image. However, since the quantity of each sign code is 1[bit/pel], the efficient reduction method for sign code is required. In this paper, more important area on the 2-D spatial domain for the sign code is determined to increase the image quality of the reconstructed image, and we have discussed the efficient preservation method of the sign code using the parameters in the obtained important area.
    Download PDF (11903K)
  • Ken KUREMATSU, Ko SAKAI
    Type: Article
    Session ID: HI2013-44/3DIT2013-1
    Published: February 26, 2013
    Released: September 21, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Symmetry is a significant factor for deriving the direction of figure. Although symmetry is common among both biological shapes and artificial objects, and known as a Gestalt factor, few studies have quantified its nature and analyzed the function quantitatively. We focus on symmetry in contours of natural images, and establish a geometrical index that describes the degree of symmetry (DoS) inherent in arbitrary contour. To examine whether DoS matches with the perception of symmetry, we carried out psychophysical experiments. The results showed a good agreement between the perception of symmetry and DoS where the participants showed confidence/consistency in the perception, indicating that DoS reflect the perception of symmetry and that perception of symmetry depends surely on the geometrical degree of symmetry.
    Download PDF (3055K)
  • Kenji YOKOI, Kyosuke MATSUSHITA
    Type: Article
    Session ID: HI2013-45/3DIT2013-1
    Published: February 26, 2013
    Released: September 21, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Although many studies on stereoscopic visual fatigue dedicated solely to a passively viewing condition such as watching movies, the influence of activeness (e.g., game-playing) on the fatigue is unclear. In this study, we compared the fatigue caused by active game-playing and passive game-watching under both 2D and 3D conditions. In addition, general 3D displays cannot reproduce motion parallax with head motion. It may be an unremarked cause of the fatigue. To clarify this possibility, head motion during the experiment was also recorded. We found that actively viewing reduced stereoscopic visual fatigue and head motion differed qualitatively according to viewing conditions. These results suggest strong connection between head motion and visual fatigue.
    Download PDF (552K)
  • Ichiro KURIKI
    Type: Article
    Session ID: HI2013-46/3DIT2013-1
    Published: February 26, 2013
    Released: September 21, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Chromaticity of light that appear achromatic under chromatic illuminant (achromatic point) depends on the lightness level of the sample in the lighting environment, such as Helson-Judd effect. The present study will report a concise method to represent the lightness dependence, and will discuss the color vision mechanisms that yield such characteristics.
    Download PDF (310K)
  • Takahiko FUKINUKI
    Type: Article
    Session ID: HI2013-47/3DIT2013-1
    Published: February 26, 2013
    Released: September 21, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Sigma movement is the so-called illusion, in which, if motion stimulus is given to a flickering checker pattern, motion appears to the same direction as the stimulus, and, has been studied by psychologists for long time. In this paper, the motion is analyzed in the three dimensional spatio-temporal frequency domain, as visible motion (physical phenomena) in the modified visible frequency band caused by eye pursuit.
    Download PDF (807K)
  • Naoki NISHIDA, Takeshi KOHAMA
    Type: Article
    Session ID: HI2013-48/3DIT2013-1
    Published: February 26, 2013
    Released: September 21, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    The purpose of this study is to understand the function of working memories while seeking a target during visual search tasks by using a neuron network model. A previous study indicated that the model which can reproduce gaze shifts while searching a target must contain a mechanism with which allocating attention to the visual attributes dynamically shifts (Sugimoto & Kohama, 2011) and maintaining inhibition to a past gazed position for several steps. However, the target information must set as a parameter before initiation of searching tasks, and also implementation of the inhibition of return in this model was not based on the physiological evidence. In this study, we proposed a modified version of Sugimoto & Kohama's model with considering the working memories for the visual attributes derived by PF and IT network. Furthermore, we added a network which assumed TPJ between V4 and LIP to memorize the locus of past gazed regions for the use of the inhibition of return. The visual search simulation results suggest that the target detection rate of our model, without giving any target information, is significantly higher than that of the previous model.
    Download PDF (3297K)
  • Mayo YAMASAKI, Takeshi KOHAMA
    Type: Article
    Session ID: HI2013-49/3DIT2013-1
    Published: February 26, 2013
    Released: September 21, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    In late years, the distribution of gaze on the media, such as Web contents, are applied to evaluate the attractiveness of users attention. Mathematical models which computes the most likely target location based on the saliency maps is frequently used to predict gaze shifts in the visual scenery. However, these models can not reproduce temporal transition of gaze positions; the history of scanned regions. The purpose of this study is to predict the gaze shifts with temporal transition in the visual search tasks. Our model reproduces gaze shifts for searching targets by using conditional probability considering scanned logs of gaze positions. The probability map is calculated from the saliency distribution. To test the performances of our model, we compared the efficiency of target detection between our model and Itti & Koch's original model in some visual search simulations. The results show that the proposed model can shift gaze positions without applying a winner-take-all process which is needed for the original model to select gaze targets. Moreover, the proposed model improves the accuracy of target detection rate, and also it can detect targets in the fewer steps than the original model. We computed the averaged diffusivity and similarity for the use of the indexes to evaluate the scan path profile of simulated gazed positions. These indexes indicate that the gaze distributions produced by the proposed model contain the temporal transition of scanned logs of the gazed positions.
    Download PDF (9909K)
  • Katsuya YANO, Takeshi KOHAMA
    Type: Article
    Session ID: HI2013-50/3DIT2013-1
    Published: February 26, 2013
    Released: September 21, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Previous saliency-based visual attention models dealt with the mechanism underlying attentional shifts in the early visual system. Such models can predict the distribution of gaze movements on the basis of saliency maps (Itti&Koch, 2000). Saliency maps are calculated by weighting to the distribution of color and the orientation of input images. Since predicted gaze positions using the previous saliency map models are deterministically selected by tracing the peaks of saliency distribution, the simulated trajectories of gaze positions show non-contextual and unnatural patterns. In this study, we purposed a mathematical model to reproduce more realistic gaze distributions. We have carried out an extension of the saliency map taking into account the spatio-temporal characteristics of attention for the visual attributes, and also image projection from retina to V1; cortical magnification factor. To evaluate the performance of our model, we conducted visual search task simulations. The results indicate that the target detection rate of our model are higher than that of Itti & Koch's model and simulated gaze distributions show contextual patterns while searching for the target.
    Download PDF (9879K)
  • Kodai TSUDA, Takeshi KOHAMA
    Type: Article
    Session ID: HI2013-51/3DIT2013-1
    Published: February 26, 2013
    Released: September 21, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Saliency map models are mathematical expressions of the characteristics of bottom-up attention processing (Itti & Koch, 2000). The saliency maps are calculated by weighting to the distribution of color and the orientation of input images on the bases of the physiological knowledge of early visual system. In late years, saliency-based models were modified to predict gaze positions for moving objects. However, these models have not been taken into the neural response characteristics of motion detection area in the brain, such as the area MT. The purpose of this study is to calculate the saliency maps for the motion components in moving images by reproducing the properties of the area MT neurons' receptive fields involved in motion perception. To evaluate the performance of proposed model, we conducted simulation experiments for the motion pop-out stimulus. The simulation results suggest that the proposed model can pick up the pop-out region more sensitively than the previous model. Furthermore, the simulation results for the movies of natural scenery, our model persistently responds to the region in which certain motion contrasts are contained whereas the previous model responds intermittently.
    Download PDF (11315K)
  • Yoshihiro ITAMIYA, Hiroko MITARAI, Atsuo YOSHITAKA
    Type: Article
    Session ID: HI2013-52/3DIT2013-2
    Published: February 26, 2013
    Released: September 21, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    In order to take "good" photos, firstly, it is necessary to clarify the subject matter. From preliminary survey, the results revealed that the professional photographers apply 1.7 types of composition on average; they tend to apply multiple types of compositions, such as triangular, diagonal and contrasting compositions, in one photo. The proposed method considers co-occurrence of recognized compositions and candidate proposing compositions. The author proposes a novel photo shooting method which suggests one or more types of compositions and superimposes the suggestions on the photo being taken. semantic differential method revealed that there is some effect of composition.
    Download PDF (3646K)
  • Mamoru IWABUCHI, Yang GUANG, Kenryu NAKAMURA
    Type: Article
    Session ID: HI2013-53/3DIT2013-2
    Published: February 26, 2013
    Released: September 21, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    In this study, a new software was developed to support people with severe physical disabilities to enable activities of daily living using Kinect for Windows. The software is called OAK (Observation and Access with Kinect). OAK enables switch operation for the user using the features of Kinect; human motion tracking and facial recognition. OAK can also keep the motion logs in a picture and a video format. This helps the user and their care givers to observe/understand the user's body motions.
    Download PDF (4066K)
  • Sho ENDO, Takeshi KOHAMA, Daisuke NOGUCHI
    Type: Article
    Session ID: HI2013-54/3DIT2013-2
    Published: February 26, 2013
    Released: September 21, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Recent studies show that the mechanism responsible for the occurrence of microsaccades, which are small involuntary shifts in the eye-gaze position, is related to the visual attention system. These studies report that the microsaccade rate increases with shift of attention allocation. In contrast, other studies show that the occurrence of microsaccades is inhibited when visual attention is intensely applied on a fixed target. It has not yet been established as to which of these conclusions is correct. In this study, we hypothesized that microsaccade rate decreases with the degree of attentional concentration. Subjects performed RSVP tasks with instructions to maintain their fixation on the alphabet characters which were displayed on the center of a CRT monitor. The degree of attention engagement of the subjects was controlled by changing the contrast of target characters. Then we analyzed the relationship between the microsaccade rate and the degree of attention engagement. The results show that the microsaccade rate was suppressed from just before the target objects were displayed, and increased after the target was extinguished. When higher concentration was required, this inhibition of microsaccade occurrence was prolonged. These results suggest that the microsaccade rate is inhibited with the concentration of visual attention in the foveal region.
    Download PDF (4210K)
  • Hiroshi Kasai, Keisuke OGAKI, Yusuke SUGANO, Yoichi SATO
    Type: Article
    Session ID: HI2013-55/3DIT2013-2
    Published: February 26, 2013
    Released: September 21, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    The goal of this research is estimating image preference from eye movements. Our method extracts statistical features from eye gaze and use a random forest classifier for supervised classification of image preference. Through the experiments, we show the efficiency of gaze information and how image preference influences on eye movements.
    Download PDF (4339K)
  • Kazuya MATSUBARA, Satoshi SHIOIRI
    Type: Article
    Session ID: HI2013-56/3DIT2013-2
    Published: February 26, 2013
    Released: September 21, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    To create good designs for simultaneous uses television and handheld devices, the evaluation method about display arrangement is essential. We examined a response time of a task, recording head, hand and eye movement to evaluate the efficiency of arrangement. To investigate how those stats are affected by display arrangement, we compared them between the different arrangement conditions. The results revealed, in efficient condition, the response times were shortens, rates of double check head motion were increase, and head motion amplitudes were enlarged. These characteristics should be useful to evaluate the efficiency of display arrangement under the simultaneous uses of two displays.
    Download PDF (733K)
  • Type: Appendix
    Pages App1-
    Published: February 26, 2013
    Released: September 21, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Download PDF (79K)
  • Type: Appendix
    Pages App2-
    Published: February 26, 2013
    Released: September 21, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Download PDF (79K)
  • Type: Appendix
    Pages App3-
    Published: February 26, 2013
    Released: September 21, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Download PDF (79K)
feedback
Top